So, what support is needed? For the maintenance staff? This is something we should really look to. Do we provide the odl test-sets? We have to give some sort of test map, which will say which part of the software is being tested by each of those test heads. For what particular purpose can you make use of the whole test set, of the different parts of the whole test. So, once we have this sort of test map test selection criteria, I think it should work fairly well. However, the person who maintains if he has built a new model he should be able to provide further test sets.

---